Relating subsidence to land-use history using remote- sensing derived data for the Mekong delta, Vietnam
1 Department of Physical Geography, Utrecht University, The Netherlands
2 Department of Subsurface and Groundwater Systems, Deltares Research Institute, Utrecht, The Netherlands 3 Departement of Earth System Science, Stanford University, Stanford
4 US EPA Office of Research and Development, National Health and Environmental Effects Research Laboratory, Atlantic Ecology Division, Narragansett, RI, USA
* Corresponding author: P.S.J.Minderhoud@uu.nl # Presenting author
Introduction
The Vietnamese Mekong delta is subsiding due to a combination of natural and human-induced causes. Over the past several decades, large-scale anthropogenic land-use changes have taken place as a result of increased agricultural production, population growth and urbanization in the delta. Land-use changes can alter the hydrological system or increase loading of
Philip S.J. Minderhoud1,2,*, Laura Coumou1,#, L.E. Erban3,4, H. Middelkoop1, E. Stouthamer1, E.A. Addink1
Figure 2. Land-use maps of the northeastern part of the Vietnamese Mekong delta derived from Landsat 5 TM imagery of 1988, 1996, 2006 and 2009.
References
Erban, L. E., Gorelick, S. M., & Zebker, H. A. (2014). Groundwater extraction, land subsidence, and sea-level rise in the Mekong Delta, Vietnam. Environmental Research Letters, 9(8), 1–6.
Acknowledgements
The Rise and Fall project is funded by NWO-WOTRO (W 07.69.105), Deltares and TNO-Geological Survey of the Netherlands.
Conclusions
• Land use and land-use history have an indirect causal relationship with subsidence rates in the Mekong delta
• Different land-use classes are experiencing different rates of subsidence
• Highest subsidence rates were found for land-use classes in which the natural environmental conditions were most altered by human activities
Figure 1. Study area with InSAR-derived subsidence rates (2006-2010 by Erban et al., 2014) in the Vietnamese Mekong delta (VMD) in Southeast Asia.
Land-use maps
Data and approach
We created a new consistent time series of land-use maps by classifying Landsat Thematic Mapper (TM) 5 images using object-based image analysis. The land-use maps were used to create the land-use history for the period 1988-2006.
Combined with InSAR-based subsidence rate, the subsidence rate for each individual land use and land-use change trajectory was quantified. To assess the strength of the relationship between land use, land-use change and subsidence, we evaluated the ability to predict subsidence rates based solely on land-use history.
the delta surface, amplifying natural subsidence processes or creating new anthropogenic subsidence. We quantified subsidence rates for the various land-use classes and past land-use changes and evaluated the relationship strength between current land use, land-use history and subsidence by predicting subsidence rates during the measurement period solely based on land-use history.
Subsidence rate per land-use type
Figure 3. Estimated impact of subsidence drivers, and mean InSAR-based subsidence rate per land-use type (similar land-use class from 1988-2006). The estimated impact of each subsidence driver is ranked: minimal (-), low (±), moderate (+) and high (++).
Figure 4. Impact of past land-use (LU) changes on the subsidence rates in the Mekong delta. The mean subsidence rates (in mm yr-1) for areas in which LU 1 and LU 2 was unchanged during the period 1988-2006 (columns ‘Unchanged LU 1’ and ‘Unchanged LU 2’) and for areas that experienced a transition from LU 1 to LU 2 between 1996-2006 (column ‘Transition from LU 1 to LU 2 1996-2006) and 1988-1996 (column ‘Transition from LU 1 to LU 2 1988-1996), respectively <10 and 10-18 years before the measurement period (2006-2010).
The effect of land-use change on subsidence rate
Land-use change
Table 1. Area in percentages of the main land-use classes in the land-use maps of 1988, 1996 and 2006. Color coded from small (green) to large (red).
Figure 5. A) Observed InSAR-based subsidence rate, B) predicted subsidence rates for the period 2006-2010, and C) difference between observed and predicted subsidence. The predicted subsidence over the period 2006-2010 (B) is based on land-use history. 66% to 92% of the predictions fall the error range of the InSAR-based subsidence (5 to 10 mm) of the observed values, observations.